1756 THE JOURNAL OF ANTIBIOTICS DEC. 1989 CISPENTACIN, A NEW ANTIFUNGAL ANTIBIOTIC II. IN VITRO AND IN VIVO ANTIFUNGALACTIVITIES TOSHIKAZU OKI, MlNORU HlRANO, KOZO TOMATSU1, KEI-ICHI NUMATA and Hideo Kamei Bristol-Myers Research Institute, Ltd., Tokyo Research Center, 2-9-3 Shimo-meguro, Meguro-ku, Tokyo 153, Japan (Received for publication June 3, 1989) Cispentacin (( - )-(li?,2^)-2-aminocyclopentane- l -carboxylic acid) is a new antifungal antibiotic possessing potent anii-candida activity. The 50% inhibitory concentration (IC50) and IC100 values of cispentacin against clinical isolates of Candida albicans were in the ranges 6.3~12.5 and 6.3 ~ 50/jg/ml, respectively, by turbidimetric measurement in yeast nitrogen base glucose medium. No significant activity was seen against any yeasts and molds whentested by the agar dilution method using three different agar media; Knopp's agar, yeast extract - glucose -peptone agar and Sabouraud dextrose agar. This antibiotic demonstrated good therapeutic efficacy against a systemic Candida infection in mice by both parenteral and po administrations. The 50%protection dose (PD50) values after single iv and po administrations were 10 and 30 mg/kg, respectively. It was also effective in a systemic infection with Cryptococcus neoformans and in both lung and vaginal infections with C. albicans in mice. Cispentacin did not induce acute lethal toxicity at 1,000mg/kg by iv injection and 1,500mg/kg by ip and po administrations in mice. In the course of antifungal screening, cispentacin, ( - )-(17?,2S)-2-aminocyclopentane-l-carboxylic acid was isolated from the fermentation broth of Bacillus cereus strain No. L450-B2. Cispentacin is a water-soluble and amphoteric compound, and possesses excellent in vivo antifungal activity against experimental infection with Candida albicans despite its apparently low in vitro activity. In the present study, cispentacin was investigated for its in vitro activity against both bacteria and fungi, and for in vivo activity against various experimental fungal infections with C. albicans and other fungi in mice. Materials and Methods Antifungal Agents Cispentacin was prepared at Bristol-Myers Research Institute, Ltd., Tokyo1}. Amphotericin B (Fungizone, sodium deoxycholate complex) and 5-fluorocytosine were purchased from Squibb Japan Inc. and Sigma Chemical Company,respectively. The above three compoundswere used by dissolving in saline for both in vitro and in vivo experiments except for a turbidimetric method where they were dissolved in 0.1 m phosphate buffer (ph 7.0). Ketoconazole was obtained from Janssen Pharmaceuticals and used as a suspension in saline containing one drop Tween 80. In Vitro Antifungal Activity The MIC's of the antifungal agents were determined by an agar dilution method on Knopp's agar (KA), yeast extract- glucose - peptone agar (YGPA) and Sabouraud dextrose agar (SDA) media. A 3-/A of fungal suspension containing 106 cells/ml was applied to the surface of the antibiotic-containing agar plates with a multiinoculator. After incubation at 28 C for 44 hours, the lowest concentration of antibiotic 1 Deceased.
VOL. XLII NO. 12 THE JOURNAL OF ANTIBIOTICS 1757 causing virtually complete inhibition of fungal growth (MIC) was determined. Susceptibility of Candida species was determined by a turbidimetric method using micro-liquid culture as follows: Candida species were shake-cultured in YGPmedium (glucose 1.5%, peptone 1.0%, yeast extract 0.4%, K2HPO40.05% and MgSO4-7H2O0.05%, ph7.0) at 28 C for 18 hours. The culture (0.1 ml) was inoculated into a fresh YGPmedium (5ml) and incubated at 28 C for 4 hours. The fungal suspension containing 1 x 107 cells/ml was diluted 1,000 times with a fresh YNGmedium (yeast nitrogen base 0.67%, glucose 1.0% and amikacin 50u/ml, ph7.0). A volume of 50fA of the test compound solution in 0.1 m phosphate buffer (ph7.0) and 200/xl of the fungal suspension were mixed in wells of a 96-well microtiter plate and incubated without agitation at 37 C for 40 hours. The cell growth was measured by an autoreader at 620nm. The turbidity of the culture without the test compound was used as a control (0%) and fresh YNGmediumas 100% inhibition control. Fifty % inhibitory concentration (IC50) and IC100 values were estimated from the values of 8 concentrations according to the method described by Galgiani and Stevens2). In Vitro Antibacterial Activity In vitro antibacterial activity of cispentacin was tested comparatively with kanamycin by an agar dilution method using nutrient agar (NA, meat extract 0.5%, peptone 0.5%, NaCl 5% and agar 1.5%, ph 7.0) and Davis synthetic agar (DA, K2HPO4 0.7%, KH2PO4 0.3%, sodium citrate 0.05%, (NH4)2SO4 0.1 %, MgSO4 0.01 %, glucose % and agar 1.5%, ph 7.0) media. A 3-/*l of bacterial suspension containing 104 cells/ml was applied to the surface of antibiotic-containing agar plate and MIC was determined after incubation at 37 C for 40 hours. In Vivo Antifungal Activity Systemic Fungal Infections: The in vivo therapeutic efficacy of cispentacin was examined comparatively with amphotericin B, 5-fluorocytosine and ketoconazole against systemic infections with Candida albicans A9540 (106 cells/mouse), 5-fluorocytosine-resistant C. albicans YA22851 (5.5 x 105 cells/mouse), Cryptococcus neoformans IAM4514 (106 cells/mouse) and Aspergillus fumigatus IAM 2034 (107 cells/mouse) in normal mice. C. albicans and C. neoformans were cultured at 28 C for 18 and 48 hours, respectively, in YGPmedium and suspended in saline. A. fumigatus was cultured at 28 C for 7 days on YGPagar slant and the spores were suspended in saline. Male ICR mice weighing 20 to 24g were infected intravenously with approx 10 times of the median lethal doses of each fungus as shown above. Groups of 5 mice at each dose level were given with the test compounds by various dosing routes and schedules. The 50% protection dose (PD50) was calculated by the method of Litchfield and Wilcoxon3) from the survival rate recorded 20 days after the fungal challenge. Untreated animals died 7 to 1 5 days post infection. Systemic Candida tropicalis Infection in Immunocompromised Mice: Groups of 5 male ICR mice weighing 20 to 24g were treated intraperitoneally with 200mg/kg of cyclophosphamide 4 days before intravenous C. tropicalis CS-0V infection at 2.3 x 106 cells/mouse (10 times of LD50). Test compounds were administered orally to the mice once immediately after the fungal challenge. Lung C. albicans Infection: Groups of 5 male ICR mice weighing 20 to 25g were treated intraperitoneally with 200 mg/kg of cyclophosphamide 4 days before the lung infection with C. albicans A9540. The mice were anesthetized with ether and were infected with a 70-/xl of C. albicans cell suspension containing 106 cells through an intranasal route on day 0. Test compounds were administered intravenously to the mice immediately after the inoculation followed by once a day dosing on days 2 and 4. The PD50 was calculated from the survival rate recorded 20 days after the fungal challenge. Vaginal C. albicans Infection: Groups of 5 female ICR mice weighing 20 to 24g were treated subcutaneously with 0.5 mg/kg ofestradiol benzoate 3 days before and 2 days after the vaginal infection with C. albicans A9540. A 10-^1 of cell suspension of C. albicans containing 108 cells/ml was inoculated intravaginally on day 0. Test compounds were administered orally to the mice once a day on days 0 to 4. On the day 7, vaginal exudate was sampled with a thin glass rod and spread on an YGPagar plate containing 100 /^g/ml of chloramphenicol. The plate was incubated at 28 C for 2 days. Viable cells count was made and graded by the following scores:
1758 THE JOURNAL OF ANTIBIOTICS DEC. 1989 No. of cells/plate >459 146-458 47-145 15-46 0-14 Score Acute Toxicity in Mice Cispentacin was dissolved in saline and administered once to groups of 5 male ICR mice at the dosed volume of 0.1 ml per 10g body weight by the iv, ip and po routes. The animals were observed daily for 10 days for physical and behavioral signs of toxicity. Results In Vitro Antifungal and Antibacterial Activities MICs for cispentacin and amphotericin B against 14 different fungi and yeasts are shown in Table 1. Only partial inhibition of the cell growth of C. albicans and Trichophyton mentagrophytes was observed at 50 and 200/xg/ml of cispentacin on YGPAmedium. No growth inhibition was seen against the other fungi on this agar medium. However, by a turbidimetric method, cispentacin exhibited in vitro anti-candida activity with the IC50 values of 6.3 ~ 12.5/ig/ml against various strains of C. albicans, 100/^g/ml against C. tropicalis and 3.1~6.3^g/ml against Candida krusei as shown in Table 2. The IC100 values were 6.3 ~50//g/ml against C. albicans, > 100/ig/ml against C. tropicalis and 6.3~ 12.5//g/ml against C. krusei. Cispentacin showed weak activity only against Bacillus subtilis PCI 219 at 25 /zg/ml on Davis synthetic agar medium. In Vivo Antifungal Activity against Systemic Fungal Infections in Mice The therapeutic efficacy of cispentacin against systemic fungal infections in mice was comparatively Table 1. Antifungal spectrum of cispentacin and amphotericin B by an agar dilution MIC Og/ml) Test organism Cispentacin Candida albicans IAM 4888 C. albicans A9540 Cryptococcus neoformans D 49 C. neoformans IAM 4514 Aspergillus fumigatus IAM 2530 A. fumigatus IAM 2034 y4../7avms NRRL 484 Fusarium moniliforme A2284 Piricularia oryzae D9 1 Trichophyton mentagrophytes D 1 55 r. mentagrophytes No. 4329 Sporothrix schenckii IFO 8158 Petriellidium boydii IFO 8078 Mucor spinosus IFO 5317 KA ( ): Concentration showing partial inhibition. ND: Not determined. KA: ph4.9, YGPA: ph6.6, SDA: ph7.0. YGPA >200(50) >400 >200(50) >400 > 200 >400 > 200 >400 >200 >400 >200(50) >400 >200 >400 KA method. Amphotericin B YGPA 12.5 () 25 3.1 () () () 50 3.1 () 50 3.1 () 100 3.1 () 50 25 (6.3) ND () ND 3.1 () ND 3.1 () 25 (3.1) 100 (3.1) 3.1 () < < 3.1 12.5 3.1 6.3 3.1 50 (3.1)
VOL. XLII NO. 12 THE JOURNAL OF ANTIBIOTICS 1759 Table 2. In vitro anti-candida activity of cispentacin, amphotericin B and 5-fluorocytosine by a turbidimetric method. Inhibition concentration (fig/ml) Cispentacin Amphotericin B 5-Fluorocytosine IC, ic10 IC, ICi ic, ic, Candida albicans IAM 4888 C. albicans A9540 C. albicans YA15049 C. albicans YA15051 C. a/frcfl/w YA22802 C. fl/foaww YA22804 C. albicans YA22851 C. a/toca/w YA22853 C, albicans YA22578 C. a/wcfl«jya258 10 C. tropicalis CS-07 C. ytrw5e/ IAM 4489 C. krusei YA15052 12.5 25 6.3 6.3 6.3 6.3 12.5 50 12.5 50 100 > 100 3.1 6.3 0.4 0.4 0.4 Table 3. In vivo antifungal activity ofcispentacin and other antifungal agents against systemic Candida infections in mice (n=5). PD50 (mg/kg/dose)a Compound Cispentacin Amphotericin B 5-Fluorocytosine Ketoconazole C. albicans A9540 Dosing route and schedule0 po Singleiv Bidx2po Qdx6iv Bidx2im Singlepo Bidx2po 10 1.9 8.8 30 0.42 0.14 0.15 >4.0 60 44 NT 70 > 100 9.0 4.9 0.60 0.65 > 200 > 200 56 NT C. albicans0 C. tropicalis YA2285 1 CS-07 Dosing Dosing route and route and schedule0 schedule0 81 0.78 75 NT b 5-Fluorocytosine-resistant strain. c The first dosing was done immediately after the fungal challenge: Qdx6, once daily for 6 days; Bidx2, twice a day for 2 days. NT: Not tested. Inoculum size: 10LD50. examined with amphotericin B, 5-fluorocytosine and ketoconazole. Anti-Candida activity was determined by various dosing routes and schedules as summarized in Table 3. Against the systemic C. albicans A9540 infection, cispentacin was highly active by iv, intramuscular and po administrations. The PD50 values of orally administered cispentacin was 9.0mg/kg by twice a day dosing for 2 days and it was approximately 6 times more potent than ketoconazole. By single po administration, cispentacin demonstrated significant anti-candida activity showing a PD50 value of 30mg/kg, while amphotericin B, 5-fluorocytosine and ketoconazole was inactive at 4.0, 100 and loomg/kg, respectively. Against 5-fluorocytosine-resistant C. albicans YA22851, cispentacin was remarkably effective by twice a day dosing for 2 days. Its PD50 value was 4.9mg/kg, approximately two times more potent than that against C. albicans A9540 by the
1760 THE JOURNAL OF ANTIBIOTICS DEC. 1989 Table 4. In vivo antifungal activity of cispentacin and other antifungal agents against systemic Cryptococcus or Aspergillus infection in mice (n = 5). Compound 42 >50 Cispen tacin Amphotericin B 0.36 0.35 5-Fluorocytosine > 100 > 100 Ketoconazole > 50 NT PD50 (mg/kg/dose)a C. neoformans IAM 4514 Dosing route and schedule13 Qdx6iv Singleiv Qdx6iv Bidx2im Singlepo Bidx5no >50 >200 0.33 3.6 > 100 >200 a PD50 value was calculated on day 20 after the fungal challenge. b The first dosing was done immediately after the fungal challenge: Qdx6, oncedailyfor6 days; Bidx2 or x5, twice aday for2days or 5 days. NT: Not tested. Inoculum size: 10LD50. A. fumigatus IAM 2034 Dosing route and schedule*3 36 1.4 8 >200 > 100 > 100 NT same dosing schedule. This antibiotic also exhibited significant therapeutic efficacy against C. tropicalis infection in immunocompromisedmice by a single po administration. As summarized in Table 4, cispentacin was inactive against systemic A. fumigatus infection but was active against systemic C. neoformans infection by once daily administration for 6 days and twice a day for 5 days with the PD50 values of 42 and 36 mg/kg, respectively. In Vivo Antifungal Activity against Lung and Vaginal Infections with C. albicans in Mice As shown in Fig. 1, intravenously administered cispentacin or amphotericin B significantly increased the survival rate of mice with lung candidiasis, their PD50 values were 28 and 0.7mg/kg, respectively. As shown in Table 5, orally administered cispentacin Table 5. Antifungal activity of cispentacin and other antifungal agents against vaginal Candida albicans A9540 infection in mice3 («=5). Compound Doseb Lesionscore PE>50 (mg/kg/ (mean (mg/kg/ dose) ± SE) dose) Vehicle control - 2.8 +0 - Cispentacin 100 0.6 + 7c 76 50 2.6+0 25 2.8+0.32 Amphotericin B 5 +0.37c 0.19 0.13 2.2+2 5-Fluorocytosine 200 2.2+0.49 >200 100 3.0+0.66 a Female ICR mice (20~24g) were treated subcutaneously with 0.5mg/kg of estradiol benzoate 3 days before and 2 days after the fungal challenge. b Test compoundswere given to mice orally once daily for 5 days. c Significantly different from control, P<0.0\. Inoculum size: 106 cells/mouse. demonstrated significant therapeutic efficacy with the PD50 value of 76mg/kg against Candida vaginal infection. Acute Toxicity in Mice No lethal toxicity was observed in mice given cispentacin at 1,000mg/kg intravenously or 1,500 mg/kg intraperitoneally and orally. Whencispentacin was administered at 1,000mg/kg orally or 1,500mg/kg intraperitoneally, muscle relaxation was observed in all mice 30 minutes after administration, and this continued for 5 days. By iv administration, the above effect was not observed even at 1,000mg/kg.
VOL. XLII NO. 12 THE JOURNAL OF ANTIBIOTICS 1761 Fig. 1. Antifungal activity of cispentacin and other antifungal agents against lung Candida albicans A9540 infection in immunocompromisedmice3. Male ICR mice (20~24 g) were injected intraperitoneally with 200mg/kg cyclophosphamide 4 days prior to fungal infection. Drugs were intravenously injected immediately after the fungal challenge followed by once a day dosing on days 2 and 4. Discussion Cispentacin is a new antibiotic produced by Bacillus cereus strain No. L450-B2 exhibiting potent in vivo anti-candida activity. In vitro antifungal activity was seen against a number of C. albicans isolates by turbidimetric analysis in YNGmedium, but only partial inhibition of the cell growth of C. albicans at 50 ^g/ml and no antibacterial activity at 100 ^g/ml was seen by agar dilution methods. In a murine systemic infection with C. albicans, cispentacin demonstrated potent therapeutic efficacy after iv, intramuscular and oral administrations. In particular, cispentacin was effective by single iv and oral dosing with the PD50 values of 10 and 30 mg/kg, respectively, suggesting that this antibiotic is well absorbed through GI tracts and has a long biological half-life. The preliminary pharmacokinetic experiments in mice demonstrated that cispentacin has high area under the curve (AUC) and is retained in the various tissues at a fairly high level along with a high blood level. Furthermore, cispentacin showed significant efficacy in experimental murine lung and vaginal candidiases and in a systemic cryptococcosis, but not in a systemic aspergillosis. Taking into consideration these paradoxical in vitro and in vivo results, microbiological and biochemical studies are being performed to determine the mechanismof action of this antibiotic. In the acute toxicity experiment in mice, cispentacin did not show any lethality at 1,000mg/kg by the
1762 THE JOURNAL OF ANTIBIOTICS DEC. 1989 iv route and at 1,500mg/kg by ip and oral route of administration. Reversible muscle relaxation was observed at higher ip and po doses. Cispentacin is a simple amino acid-type of antibiotic having the chemical structure of ( - )-(li?,2s)-2-aminocyclopentane-l-carboxylic acid. 2-Aminocyclopentane-l-carboxylic acid was synthesized by Plieninger and Schneider in 19594) and a series of aminocyclopentanecarboxylic acids and aminocyclohexanecarboxylic acids have been synthesized to determine some aspects of the spatial dimensions of GABAreceptors in the frog spiral cord5) and in the rat hippocampus6), but no other biological activities have been reported. Yamaguchi et al?) recently isolated rather simple amino acid antifungal antibiotic, (*S)-2-amino-5-hydroxy-4-oxopentanoic acid (RI-33 1), from a strain of Streptomyces species, which was active against yeasts but inactive against aspergilli in vitro. Orally administered RI-331 exhibited in vivo activity against murine systemic candidiasis with low acute toxicity. Acknowledgment The authors wish to thank Dr. H. Kawaguchi of this Institute for valuable discussion throughout the present study and K. Shiina for the supply of cispentacin. The excellent technical assistance of M. Yamanaka, Y. Obi, M. Mori, H. Fujimura and A. Ohta is great fully acknowledged. References 1) Konishi, M.; M. Nishio, K. Saitoh, T. Miyaki, T. Oki & H. Kawaguchi: Cispentacin, a new antifungal antibiotic. I. Production, isolation, physico-chemical properties and structure. J. Antibiotics 42: 1749-1755, 1989 2) Galgiani, J. N. & D. A. Stevens: Antimicrobial susceptibility testing of yeasts: a turbidimetric technique independent of inoculum size. Antimicrob. Agents Chemother. 10: 721-726, 1976 3) Litchfield, J. T. & F. Wilcoxon: A simplified method of evaluating dose-effect experiments. J. Pharmacol. Exp. Ther. 96: 99-113, 1949 4) Plieninger, H. & K. Schneider: Die Anlagerung von Ammoniak an z^-tetrahydrobenoesaure and Cyclopenten-(l)-carbonsaure-(l) und die eindeutige sterische Zuordnung der entstehenden /?-Aminosauren. Chem. Ber. 92: 1594-1599, 1959 5) Nicoll, R. A.: The effect of conformationally restricted amino acid analogues on the frog spinal cord in vitro. Br. J. Pharmacol. 59: 303-309, 1977 6) Segal, M.; K. Sims & E. Smissman: Characterization of an inhibitory receptor in rat hippocampus: a microiontophoretic study using conformationally restricted amino acid analogues. Br. J. Pharmacol. 54: 181-188, 1975 7) Yamaguchi, H.; K. Uchida, T. Hiratani, T. Nagase, N. Watanabe & S. Omura: RI-331, a new antifungal antibiotic. Ann. N. Y. Acad. Sci. 544: 188-189, 1989